11-09, 14:45–15:30 (America/New_York), Winter Garden (5th floor)
You might have heard that "If you are not a liberal at 25, you have no heart. If you are not a conservative at 35, you have no brain." But is it true? According to data from the General Social Survey, yes… and no. People over 30 are more likely to say they are conservative, and older people hold more conservative views. But people become more liberal as they get older, on average, and the population has been getting consistently more liberal for the last 50 years. If that sounds paradoxical, you should come to this talk, where I will present the tools we need to understand these results: age-period-cohort analysis, item response theory, Simpson's paradox, and the Overton window.
Outline:
* Young liberals, old conservatives?
* Political labels by age
* Political labels by cohort
* Defining the issues: 15 questions
* Political views over time: the Overton window
* Political views by cohort: Simpson's paradox
* Age-period-cohort analysis
* Polarization in the 2010s
* Explaining the paradoxes
Methods I'll present include resampling, local regression with LOWESS, logistic regression, and Bayesian statistics, all using NumPy, SciPy, and Pandas.
No previous knowledge expected
Allen Downey is a Staff Scientist at DrivenData and Professor Emeritus at Olin College.
He is the author of several textbooks -- including Think Python, Think Bayes, and Elements of Data Science -- and "Probably Overthinking It", a blog about data science and Bayesian statistics. He received a Ph.D. in computer science from U.C. Berkeley and Bachelor's and Master's degrees from MIT.